如何使用财务报表API做基本面分析

很多人使用行情 API 时只关注实时价格、K 线、盘口和成交明细。但真正做中长期选股、财务质量分析、估值判断、股息组合监控时价格数据是不够的。价格只是反映了市场如何给这家公司定价但财务报表能真实揭露这家公司到底赚不赚钱、现金流好不好、负债压力大不大、分红是否可持续。这一期我们将基于财务报表接口通过代码快速帮我们分析一家公司的基本面。一、财报接口数据总览下面是Infoway API的财务报表接口能查询到的数据分析方向接口路径适合做什么利润表/common/basic/financial/income_statement收入、成本、利润、净利润增速分析收入明细/common/basic/financial/revenue业务结构、地区收入结构、产品线变化现金流量表/common/basic/financial/cash_flow经营现金流、自由现金流、利润含金量资产负债表/common/basic/financial/balance_sheet总资产、负债、权益、现金、债务分析估值与统计指标/common/basic/financial/statisticsPE、PB、ROE、市值、净利率等指标股息指标/common/basic/financial/dividendDPS、股息率、派息率、分红趋势历史派息记录/common/basic/financial/dividend_payout除权日、登记日、派息日、派息金额业绩预测/common/basic/financial/earningsEPS 和营收实际值与预期值对比二、封装一个财务报表客户端先安装依赖pip install requests pandas matplotlib python-dotenv把 API Key 放到环境变量里不要直接写死在代码中。创建 .env 文件INFOWAY_API_KEY你的_api_key最后写一个通用请求函数import os import time import requests import pandas as pd import matplotlib.pyplot as plt from dotenv import load_dotenv load_dotenv() API_KEY os.getenv(INFOWAY_API_KEY) BASE_URL https://data.infoway.io if not API_KEY: raise RuntimeError(请先在 .env 文件中设置 INFOWAY_API_KEY) class InfowayFinancialClient: def __init__(self, api_key: str, base_url: str BASE_URL, timeout: int 20): self.api_key api_key self.base_url base_url.rstrip(/) self.timeout timeout self.headers { User-Agent: Mozilla/5.0, Accept: application/json, apiKey: api_key, } def get(self, path: str, params: dict | None None) - dict: url f{self.base_url}{path} resp requests.get( url, headersself.headers, paramsparams, timeoutself.timeout, ) if resp.status_code ! 200: raise RuntimeError( fHTTP 请求失败: {resp.status_code}, body{resp.text[:500]} ) data resp.json() if data.get(ret) ! 200: raise RuntimeError(fAPI 返回异常: {data}) return data def financial_statement( self, endpoint: str, symbol: str, stock_type: str, period_type: str | None None, ) - pd.DataFrame: 通用财报接口。 endpoint 示例 - income_statement - revenue - cash_flow - balance_sheet - statistics - dividend - earnings params { symbol: symbol, type: stock_type, } if period_type: params[period_type] period_type raw self.get(f/common/basic/financial/{endpoint}, paramsparams) rows raw.get(data, []) return pd.DataFrame(rows) def dividend_payout(self, symbol: str, stock_type: str) - pd.DataFrame: 历史派息记录接口。 官方说明该接口没有 period_type 参数。 params { symbol: symbol, type: stock_type, } raw self.get(/common/basic/financial/dividend_payout, paramsparams) rows raw.get(data, []) df pd.DataFrame(rows) for col in [exDate, recordDate, paymentDate]: if col in df.columns: # 官方示例是 Unix 时间戳。这里兼容秒级和毫秒级。 df[col] pd.to_numeric(df[col], errorscoerce) unit ms if df[col].dropna().gt(10**12).any() else s df[col] pd.to_datetime(df[col], unitunit, errorscoerce) return df client InfowayFinancialClient(API_KEY)三、基本面分析1. 净利润增速追踪器如果我们想了解公司最近几个季度的盈利动能是在增强还是在减弱可以用利润表接口拉取季度数据筛选net_income然后计算环比和同比。def extract_item(df: pd.DataFrame, item_id: str) - pd.DataFrame: 从长表中提取某个财务科目。 if df.empty: return pd.DataFrame() result df[df[itemId] item_id].copy() if result.empty: return result result[periodDate] pd.to_datetime(result[periodDate]) result[itemValue] pd.to_numeric(result[itemValue], errorscoerce) result result.sort_values(periodDate) return result income client.financial_statement( endpointincome_statement, symbolAAPL.US, stock_typeSTOCK_US, period_typefq, ) net_income extract_item(income, net_income) net_income[qoq_growth] net_income[itemValue].pct_change() net_income[yoy_growth] net_income[itemValue].pct_change(4) print( net_income[ [symbol, periodDate, itemValue, qoq_growth, yoy_growth] ].tail(8) )将结果可视化plt.figure(figsize(10, 5)) plt.plot(net_income[periodDate], net_income[itemValue] / 1e9, markero) plt.title(Net Income Trend) plt.xlabel(Period) plt.ylabel(Net Income) plt.xticks(rotation45) plt.tight_layout() plt.show()进一步可以画同比增速plt.figure(figsize(10, 5)) plt.bar(net_income[periodDate], net_income[yoy_growth] * 100) plt.title(Net Income YoY Growth) plt.xlabel(Period) plt.ylabel(YoY Growth (%)) plt.xticks(rotation45) plt.tight_layout() plt.show()这个分析很适合做成一个盈利动能仪表盘。如果净利润连续几个季度同比改善说明利润趋势可能正在修复如果收入增长但净利润下滑就要进一步看毛利率、费用率或一次性损益。2利润含金量分析利润表上的净利润不等于真金白银。一个公司可能净利润很好看但经营现金流很差这时就要警惕应收账款、存货、资本化处理或利润质量问题。我们可以用经营现金流 / 净利润 自由现金流 / 净利润 自由现金流率 自由现金流 / 营业收入代码如下cash_flow client.financial_statement( endpointcash_flow, symbolAAPL.US, stock_typeSTOCK_US, period_typefq, ) income client.financial_statement( endpointincome_statement, symbolAAPL.US, stock_typeSTOCK_US, period_typefq, ) net_income extract_item(income, net_income) revenue extract_item(income, total_revenue) free_cash_flow extract_item(cash_flow, free_cash_flow) operating_cf extract_item(cash_flow, cf_oper) quality ( net_income[[periodDate, itemValue]] .rename(columns{itemValue: net_income}) .merge( revenue[[periodDate, itemValue]].rename(columns{itemValue: revenue}), onperiodDate, howleft, ) .merge( free_cash_flow[[periodDate, itemValue]].rename( columns{itemValue: free_cash_flow} ), onperiodDate, howleft, ) .merge( operating_cf[[periodDate, itemValue]].rename( columns{itemValue: operating_cash_flow} ), onperiodDate, howleft, ) ) quality[ocf_to_net_income] quality[operating_cash_flow] / quality[net_income] quality[fcf_to_net_income] quality[free_cash_flow] / quality[net_income] quality[fcf_margin] quality[free_cash_flow] / quality[revenue] print(quality.tail(8))同样将结果可视化:plt.figure(figsize(10, 5)) plt.plot( quality[periodDate], quality[ocf_to_net_income], markero, labelOperating CF / Net Income, ) plt.plot( quality[periodDate], quality[fcf_to_net_income], markero, labelFCF / Net Income, ) plt.axhline(1.0, linestyle--) plt.title(Profit Quality) plt.xlabel(Period) plt.ylabel(Ratio) plt.legend() plt.xticks(rotation45) plt.tight_layout() plt.show()一般来说长期经营现金流明显低于净利润要谨慎如果自由现金流长期为负但公司又持续扩张也要进一步区分是成长性投入还是经营质量恶化。3资产负债表健康度分析资产负债表可以回答公司是不是靠高负债撑起来的现金是否充足权益是否稳定增长我们可以计算资产负债率 总负债 / 总资产 净债务率 净债务 / 股东权益 现金债务覆盖率 现金及等价物 / 总债务代码如下:balance client.financial_statement( endpointbalance_sheet, symbol00700.HK, stock_typeSTOCK_HK, period_typefq, ) total_assets extract_item(balance, total_assets) total_liabilities extract_item(balance, total_liabilities) total_equity extract_item(balance, total_equity) cash extract_item(balance, cash_and_equiv) total_debt extract_item(balance, total_debt) net_debt extract_item(balance, net_debt) bs ( total_assets[[periodDate, itemValue]] .rename(columns{itemValue: total_assets}) .merge( total_liabilities[[periodDate, itemValue]].rename( columns{itemValue: total_liabilities} ), onperiodDate, howleft, ) .merge( total_equity[[periodDate, itemValue]].rename( columns{itemValue: total_equity} ), onperiodDate, howleft, ) .merge( cash[[periodDate, itemValue]].rename( columns{itemValue: cash_and_equiv} ), onperiodDate, howleft, ) .merge( total_debt[[periodDate, itemValue]].rename( columns{itemValue: total_debt} ), onperiodDate, howleft, ) .merge( net_debt[[periodDate, itemValue]].rename( columns{itemValue: net_debt} ), onperiodDate, howleft, ) ) bs[debt_asset_ratio] bs[total_liabilities] / bs[total_assets] bs[net_debt_to_equity] bs[net_debt] / bs[total_equity] bs[cash_to_debt] bs[cash_and_equiv] / bs[total_debt] print(bs.tail(8))可视化plt.figure(figsize(10, 5)) plt.plot(bs[periodDate], bs[debt_asset_ratio] * 100, markero) plt.title(Debt Asset Ratio) plt.xlabel(Period) plt.ylabel(Debt / Assets (%)) plt.xticks(rotation45) plt.tight_layout() plt.show()这个分析适合用来筛查“高杠杆扩张”风险。比如收入和利润增长很快但资产负债率同步快速上升且经营现金流没有改善就可能是比较危险的增长。4收入结构可视化收入明细接口非常适合分析公司业务结构。比如苹果可以拆成 iPhone、Mac、iPad、Services、Wearables也可以按地区拆成 Americas、Europe、Greater China 等。revenue_detail client.financial_statement( endpointrevenue, symbolAAPL.US, stock_typeSTOCK_US, period_typefy, ) print(revenue_detail.head()) print(revenue_detail[itemGroup].dropna().unique())筛选业务收入business_rev revenue_detail[ revenue_detail[itemGroup] Revenue by business ].copy() business_rev[periodDate] business_rev[periodDate].astype(str) business_rev[itemValue] pd.to_numeric(business_rev[itemValue], errorscoerce) pivot business_rev.pivot_table( indexperiodDate, columnsitemName, valuesitemValue, aggfuncsum, ).sort_index() print(pivot.tail())画堆叠面积图plt.figure(figsize(12, 6)) plt.stackplot( pivot.index, [pivot[col] / 1e9 for col in pivot.columns], labelspivot.columns, ) plt.title(Revenue by Business) plt.xlabel(Fiscal Year) plt.ylabel(Revenue) plt.legend(locupper left) plt.xticks(rotation45) plt.tight_layout() plt.show()这个分析比单纯看总收入有意思得多。因为很多公司的估值重估并不是来自总收入增长而是来自收入结构变化。比如硬件公司服务化、周期股产品结构升级、传统公司高毛利业务占比提升都可能改变市场给它的估值方式。5价值股筛选器估值与统计指标接口可以用来批量筛选股票。比如我们想找PE 15PB 2ROE 15%净利率 10%先写一个函数把 statistics 中的关键指标取出来def safe_float(x): if x is None: return None if isinstance(x, str) and x.strip() in {, -}: return None try: return float(x) except Exception: return None def latest_statistics(symbol: str, stock_type: str) - dict: df client.financial_statement( endpointstatistics, symbolsymbol, stock_typestock_type, period_typefq, ) if df.empty: return {symbol: symbol} # 优先用 currentValue没有就用 itemValue latest_date df[periodDate].iloc[0] latest df[df[periodDate] latest_date].copy() result { symbol: symbol, periodDate: latest_date, } for _, row in latest.iterrows(): item_id row.get(itemId) value row.get(currentValue) if value is None or value : value row.get(itemValue) result[item_id] safe_float(value) return result批量筛选symbols [ (AAPL.US, STOCK_US), (MSFT.US, STOCK_US), (TSLA.US, STOCK_US), (00700.HK, STOCK_HK), (000001.SZ, STOCK_CN), ] rows [] for symbol, stock_type in symbols: try: rows.append(latest_statistics(symbol, stock_type)) time.sleep(0.2) except Exception as e: print(f{symbol} 查询失败: {e}) factor_df pd.DataFrame(rows) print(factor_df.head())不同接口中的itemId命名可能会随市场和数据源略有差异。你可以先打印所有字段print(factor_df.columns.tolist())